• Cultural analytics — Using data science to study cultural production, circulation, and reception (texts, images, media, institutions).
  Example: Measuring how Latin American women novelists are cited across decades and languages.
• Property/knowledge graph — A graph where nodes/edges have attributes (property graph) and often typed relations consistent with a schema/ontology (knowledge graph).
  Example: Nodes = authors with attributes (gender, region); edges = “cites”, “translated_to”.
• Node / Edge / Attribute — Nodes are entities (author, film, museum); edges are relations (citation, collaboration); attributes add context (year, genre, language).
  Example: A “Borges → Kafka” edge with relation="cites" and year=1954.
• Bipartite graph — Two types of nodes with edges only across types (e.g., people↔films).
  Example: People–film participation; later project to a people↔people graph weighted by shared films.
• Multiplex / multilayer graph — Multiple relation types or layers on the same set of nodes.
  Example: Author graph with layers for “cites”, “coauthors”, and “mentions”, analyzed together.
• Projection (of bipartite graphs) — Building a one-mode graph from a two-mode graph by connecting nodes that co-occur.
  Example: Two artists connected if they co-exhibited in at least one show; edge weight = number of shared shows.

Network structure & metrics

• Degree / Weighted degree — Number of ties (or sum of weights) a node has; a proxy for activity/visibility.
  Example: Which director has the most repeated collaborations in a decade?
• PageRank / Eigenvector centrality — Influence via connections to influential neighbors; good for “canon” detection.
  Example: Ranking philosophers by incoming citations from already central philosophers.
• Betweenness centrality — Fraction of shortest paths that pass through a node; highlights bridges/gatekeepers.
  Example: Identifying the curator who links otherwise separate museum circuits.
• Closeness centrality — How close a node is, on average, to all others (reachability/efficiency).
  Example: Which festival gives an artist the shortest path to most other festivals?
• Clustering coefficient — How inter-connected a node’s neighbors are (local cohesion).
  Example: Tight artistic scenes vs. loosely connected collectives.
• Burt’s structural holes / Constraint — Low constraint = brokerage power across otherwise disconnected groups.
  Example: A translator connecting minoritized authors to mainstream publishers.
• Community detection (Louvain/Leiden) — Partitioning into densely connected groups (subcultures, schools, circuits).
  Example: Distinct subgenres in a sampling/influence network in hip-hop.
• Assortativity / Homophily — Do ties cluster by attributes (gender, region, language)? +1 segregated; 0 mixed; −1 cross-mixing.
  Example: Are women screenwriters collaborating mostly with women?
• Exposure / Neighborhood composition — Share of neighbors with a given attribute (who “sees” whom).
  Example: Average % of non-EU artists in a museum’s co-exhibition neighborhood.
• Centrality mass by group — Share of total centrality a group holds (fairness diagnostic).
  Example: % of PageRank held by Global South journals in a citation graph.
• Motifs / k-core / articulation points — Small recurring subgraphs; nested cores of cohesion; nodes whose removal disconnects the graph.
  Example: Triadic motifs in film crews; articulation festivals whose removal splits the touring circuit.

Dynamics & inference

• Temporal (dynamic) networks — Time-sliced graphs to track change.
  Example: How gender assortativity in film collaborations changes every three years.
• Diffusion / Cascades — Tracing how memes, styles, or ideas spread over time.
  Example: Pathways by which a dance trend crosses language communities on TikTok.
• Link prediction — Estimating likely future edges from topology/embeddings.
  Example: Predicting future co-exhibitions between artists.
• Influence maximization / backbone extraction — Choosing seed nodes to maximize spread; pruning to the most informative edges.
  Example: Selecting a few festivals to maximize international exposure for emerging artists.

Content & graph hybrids

• Embedding-based similarity graph — Connect items by semantic similarity (text/image/audio embeddings).
  Example: A book-to-book similarity network that you cluster and then audit for diversity.
• Fairness dashboard (graph) — A compact set of equity metrics over time: assortativity, exposure, centrality mass, reciprocity.
  Example: Quarterly monitoring of a museum’s acquisition and exhibition networks.

Ethics & critical frameworks (with graph-friendly operationalizations)

• Epistemic violence — Harm enacted when knowledge from certain communities is excluded, misrepresented, or structurally discounted in canons and platforms.
  Graph example: In a Wikipedia biography network, a high asymmetry of incoming vs. outgoing links for women scholars (low visibility inlinks, many outlinks to male hubs) plus low reciprocity and peripheral positioning indicates epistemic violence.
• Decoloniality — A project to delink knowledge, methods, and institutions from colonial power structures and to center local, plural epistemologies.
  Graph example: Building region-aware, language-aware knowledge graphs where Latin American or African intellectual networks are analyzed on their own terms (not only via links to Euro-American hubs), and reporting metrics stratified by region/language with corrective sampling.
• Data feminism — Principles for making data work more equitable by examining power, context, and whose interests are served or harmed.
  Graph example: Publishing a fairness dashboard with group-wise centrality mass, exposure, and assortativity; including participatory interpretation with affected communities.
• Feminist AI — Designing AI systems that center care, accountability, and justice, resisting bias and amplifying marginalized perspectives.
  Graph example: A recommendation system on a cultural similarity graph that optimizes for relevance anddiversity/exposure of underrepresented creators, with transparent knobs and impact audits.
• Data sovereignty — The right of people/communities (often Indigenous/minoritized) to control data about them—how it’s collected, used, and shared.
  Graph example: A museum collaboration graph where contributors consent to inclusion, can set sharing bounds on node attributes (e.g., community affiliation), and where analyses respect governance protocols (e.g., opt-out removes nodes/attributes from all projections).

---

Mini “how you’d measure it” examples

• Under-exposure of a group: compute group’s share of PageRank vs. share of nodes; large gap ⇒ under-/over-representation.
• Segregation: attribute assortativity by gender/region over time; decreasing trend ⇒ improving mixing.
• Hidden bridges: flag nodes with high betweenness but low in-degree/PageRank (bridges with little visibility).
• Epistemic violence proxy (one operationalization):
  EVI≈EVI≈ (centrality gap index) + (1 − reciprocity) + (peripherality score) + (exposure deficit) — normalized per group/time slice.
  Use: compare EVI across regions/genders and track change after policy shifts.